IMPROVING FEATURE GENERALIZABILITY WITH MULTITASK LEARNING IN CLASS INCREMENTAL LEARNING

Many deep learning applications, like keyword spotting, require the incorporation of new concepts (classes) over time, referred to as Class Incremental Learning (CIL). The major challenge in CIL is catastrophic forgetting, i.e., preserve as much of the old knowledge as possible while learning new tasks. Various techniques, such as regularization, knowledge distillation, and exemplar, have been proposed to resolve this issue. However, prior works only focus on the incremental learning step, while ignoring the optimization of the base model training stage. We hypothesise that a more transferable and generalizable feature representation of the base model would be beneficial to incremental learning. In this work, we adopt multitask learning during base model training to improve the feature generalizability. Specifically, instead of training a single model with all the base classes, we decompose the base classes into multiple subsets and regard each of them as a task. These tasks are trained concurrently and a shared representation is obtained as the input to the incremental learning. We evaluate our approach on two datasets under various configurations. The results show that our approach enhances the average incremental learning accuracy by up to 5.5%, which enables more reliable and accurate keyword spotting over time.